Qwen2.5-1.5B-Open-R1-GRPO/train_results.json

{
    "total_flos": 0.0,
    "train_loss": 0.11724991016089917,
    "train_runtime": 482.6536,
    "train_samples": 16,
    "train_samples_per_second": 0.332,
    "train_steps_per_second": 0.021
}